Analisis Review Aplikasi MyPertamina Menggunakan Support Vector Machine

Dewi, Putri Ayu Permata (2026) Analisis Review Aplikasi MyPertamina Menggunakan Support Vector Machine. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043211077_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Transformasi digital telah menjadi elemen kunci dalam upaya peningkatan efisiensi dan transparansi layanan publik di Indonesia. Salah satu implementasinya adalah kebijakan penyaluran subsidi Bahan Bakar Minyak (BBM) berbasis Quick Response Code (QR Code) melalui aplikasi MyPertamina. Kebijakan ini bertujuan untuk memastikan distribusi subsidi lebih tepat sasaran. Namun, dalam pelaksanaannya aplikasi ini memicu beragam reaksi dari masyarakat. Beragam tanggapan ini terekam dalam ulasan pengguna di Google Play Store. Penelitian ini bertujuan untuk menganalisis sentimen publik terhadap aplikasi MyPertamina menggunakan pendekatan text mining dan algoritma Support Vector Machine (SVM) dengan kernel Radial Basis Function (RBF). Data yang digunakan dalam penelitian ini berupa ulasan pengguna aplikasi MyPertamina di Google Play Store. Data dikumpulkan dari versi aplikasi 4.7.0 hingga 4.7.2, yaitu versi yang dirilis pada periode pengambilan data dan dipilih karena masih berada dalam kelompok versi yang sama sehingga perubahan fitur diperkirakan bersifat minor. Total data yang berhasil dikumpulkan berjumlah 1015 ulasan, yang diambil dalam rentang waktu 30 September 2025 hingga 17 November 2025. Tahapan penelitian meliputi pengumpulan data ulasan, text pre-processing (case folding, cleaning, tokenizing, stemming, dan stopword removal), pelabelan sentimen, pembentukan model klasifikasi SVM, serta evaluasi model menggunakan confusion matrix. Hasil yang didapatkan dari analisis ini adalah proporsi sentimen negatif cenderung lebih besar daripada sentimen positif dengan persentase masing-masing sebesar 54.4% dan 45.6%. Hasil analisis klasifikasi menggunakan SVM RBF menggunakan parameter C sebesar 10 dan gamma sebesar 0.1 didapatkan bahwa model menghasilkan nilai accuracy sebesar 0.8703 dan hasil evaluasi menggunakan kurva ROC menunjukkan bahwa kurva berada jauh di atas garis diagonal dengan nilai AUC sebesar 0.8857 yang menunjukkan bahwa model SVM RBF memiliki kemampuan yang sangat baik dalam membedakan sentimen positif dan negatif.
========================================================================================================================================
Digital transformation has become a key element in efforts to improve the efficiency and transparency of public services in Indonesia. One example of this is the policy of distributing fuel subsidies based on Quick Response Codes (QR Codes) through the MyPertamina app. This policy aims to ensure that subsidies are distributed more accurately. However, the implementation of this app has sparked mixed reactions from the public. These diverse responses are recorded in user reviews on Google Play Store. This study aims to analyze public sentiment towards the MyPertamina application using a text mining approach and a Support Vector Machine (SVM) algorithm with a Radial Basis Function (RBF) kernel. The data used in this study are user reviews of the MyPertamina application on Google Play Store. The data was collected from app versions 4.7.0 to 4.7.2, which were released during the data collection period and were selected because they were still in the same version group, so feature changes were expected to be minor. A total of 1,015 reviews were collected, taken between September 30, 2025, and November 17, 2025. The research stages included collecting review data, text pre-processing (case folding, cleaning, tokenizing, stemming, and stopword removal), sentiment labeling, SVM classification model formation, and model evaluation using a confusion matrix. The results obtained from this analysis showed that the proportion of negative sentiment tended to be greater than positive sentiment, with percentages of 54.4% and 45.6%, respectively. The results of the classification analysis using SVM RBF with parameters C of 10 and gamma of 0.1 show that the model produces an accuracy value of 0.8703 and the evaluation results using the ROC curve show that the curve is well above the diagonal line with an AUC value of 0.8857, indicating that the SVM RBF model has excellent ability in distinguishing positive and negative sentiments.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Sentiment Analysis, Google Play Store, MyPertamina, Radial Basis Function, Support Vector Machines, Analisis Sentimen, Google Play Store, MyPertamina, Radial Basis Function, Support Vector Machines
Subjects:	Q Science Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Putri Ayu Permata Dewi
Date Deposited:	04 Feb 2026 01:46
Last Modified:	04 Feb 2026 01:46
URI:	http://repository.its.ac.id/id/eprint/131987

Actions (login required)

View Item